FEATURED Building PWM Controller for 4 wires PWM fan

[inVain]

okay,
for they who interested in my 0-100% PWM controller,
here's my last update (all images were showed in thumbnail, please click them to enlarge):

(top view)


(bottom view)

well it's not that neat in the end, I forgot to wire some pins and it end up with me to give them some "hard core spacer" style LOL

but here's the interesting part:

see all those red boxes,
they were IC sockets which I wrecked in purpose, I intend to use them as a socket . so anyone who would like to use different values of caps as the frequency filter will be able to swap them easily. and as for the resistors, I try to avoid that "need to turn around" pots like that happened with my previous project.

well I hope this failed project of mine can help the other's projects, and if anyone had any different methods to do this, please share how yo do it


I wont give up just yet on those 0-100% duty cyle controller,
so here's the rev.B board

(top view)

the light green box on the left is where the pots should take place

(bottom view)

with this bigger PCB spaces, we can arrange the wiring line, and hopefully; easier check up if something's wrong,
and as you can see, the red line will be intended as the 5v, the black line will be for ground line, and the yellow for the 12v line.
the alligator clips were intended to hold the wire in place, this shall make the soldering easy


and here's the video for my 2-98% taming that noisy delta


whoa, the script really does the magic
thx a lot bing

(on da 2nd thought; I really should throw that sandals before taking the cam LOL)

 

[bing]

Another great update, can't wait to see the final form of this 0-100% version kicking.

Just take your time on that second trial, just make sure you do check, re-check and check again before powering it up.

PS : To embed a youtube video box in your post, use this bbcode -> [yvid]bfuAvqG8qg0[/yvid]
The "bfuAvqG8qg0" is the suffix of the youtube link.

 

[inVain]

here's my 2nd attempt and still failed
the fan's stand still when I hooked up the RPM wire :



gonna visit the shop and buy the replacement IC, and this time breadboard is a must

and I wonder, if any of you guys know,
what is the F00 on the delta fan means?
here's the datasheet shown:

it's found on my PFC1212DE, I hooked it up to my RPM sensor on my mobo, but no RPM reading
just curious what wire is it??


and come' on guys,
where are the other projects pics....
I need more inspirations for this 0-100 controller, and even to mods the 2-98

 

[inVain]

finnaly,
got this 0-100% controller work
I got a defected LM311, that's the only sole probrem . after replacing the IC, both board are working fine

and I really need help from you guys, I've got these caps from a long abandoned project,
here how its looks:

Spoiler

the problem is, I'm not sure if these caps are 220pF or 220nF. any help are much appreciated


and after a replacing the IC, I found out that the same IC with different manufacturer can resulting a different controller character,
as you can see, I have LM311 and NE555 manufactured by the "ST" (with CHN mark) and the "Texas Instrument"

with the "ST" on the circuit, the fan was fully stop at the most minimum,
and with the "TI", my delta was still running at the most minimum pots position

 

[bing]

and I wonder, if any of you guys know,
what is the F00 on the delta fan means?
here's the datasheet shown:

it's found on my PFC1212DE, I hooked it up to my RPM sensor on my mobo, but no RPM reading
just curious what wire is it??

That F00 suffix for Delta fan usually stands for RPM reporting.

Try the fan at full speed, without using controller to see if this fix the rpm reading.

and I really need help from you guys, I've got these caps from a long abandoned project,
here how its looks:

the problem is, I'm not sure if these caps are 220pF or 220nF. any help are much appreciated

Its 22 pf !

From the look, that might be the silver mica cap, good quality and usually they're used in oscillating circuit since they're known to have good stability.

finnaly,
got this 0-100% controller work
I got a defected LM311, that's the only sole probrem . after replacing the IC, both board are working fine

Great !

Another OCfer successfully made the 0-100% version.

Another lesson learned for those who is going to build this circuit, when it comes to these dirt cheap components, buy more sets.

and after a replacing the IC, I found out that the same IC with different manufacturer can resulting a different controller character,
as you can see, I have LM311 and NE555 manufactured by the "ST" (with CHN mark) and the "Texas Instrument"

with the "ST" on the circuit, the fan was fully stop at the most minimum,
and with the "TI", my delta was still running at the most minimum pots position

Hmm.. that is strange.

Now, "WITHOUT" the fan controller, try connect the PWM cable to the Ground wire and power up, this connection is equal to pure 0% duty cycle and tell us how is that fan behaves.

As you read thru this gigantic thread, we all learned that many OCfers encountered so many different behaviours on different PWM fans.

At 0%, some are normal with minimum speed, but not stopped, some like BF's sanyo will be totally stopped, while like Bob's weird one which has the transfer function that it will run at full speed at 0% duty cycle, but strangely at approx 5% dutycycle, it will rotate at the minimum speed. Yes, its quite confusing, but all of these are made and designed for a purpose.

 

[Bobnova]

I like the 0% = 100% nidec idea, if the voltage controller eats it it'll either short to 5v, short to ground, or fail open.
If it shorts to 5v I get full fan speed and no overheating.
If it shorts to ground I get full fan speed and no overheating.
And if it shorts open I get full fan speed and no overheating.
Beautiful!

It did confuse me when I was trying to figure out if it was PWM by shorting pins to ground though!

 

[bing]

I like the 0% = 100% nidec idea, if the voltage controller eats it it'll either short to 5v, short to ground, or fail open.
If it shorts to 5v I get full fan speed and no overheating.
If it shorts to ground I get full fan speed and no overheating.
And if it shorts open I get full fan speed and no overheating.
Beautiful!

It did confuse me when I was trying to figure out if it was PWM by shorting pins to ground though!

Yep, that characteristic is designed for high reliability system.

At a really bad condition that if the fan control circuit failed either zapped/burned/toasted or etc, the fan if still has the power, must hit the max speed in order to cool the system.

Usually this kinda stuff is for industrial grade electronic control system to ensure maximum reliability.

 

[TorqueRanger]

finnaly,
got this 0-100% controller work
I got a defected LM311, that's the only sole probrem . after replacing the IC, both board are working fine

and I really need help from you guys, I've got these caps from a long abandoned project,
here how its looks:

Spoiler

the problem is, I'm not sure if these caps are 220pF or 220nF. any help are much appreciated


and after a replacing the IC, I found out that the same IC with different manufacturer can resulting a different controller character,
as you can see, I have LM311 and NE555 manufactured by the "ST" (with CHN mark) and the "Texas Instrument"

with the "ST" on the circuit, the fan was fully stop at the most minimum,
and with the "TI", my delta was still running at the most minimum pots position

Well if you look at the last letter of model of the chip it's different... So that's why your having some issues ...

The chip
TI-IM311(P)
ST-IM311

I noticed this before when i purchased the chips from mouser but was not sure if they made a difference.. One chip would be for general and next one would be for precision

 

[inVain]

That F00 suffix for Delta fan usually stands for RPM reporting.

Try the fan at full speed, without using controller to see if this fix the rpm reading.

got that...
I'll try that as soon as I got the male molex connector to connect it sraight from the PSU, I'm currently out of it

Its 22 pf !

From the look, that might be the silver mica cap, good quality and usually they're used in oscillating circuit since they're known to have good stability.

ahhhh....
I really thought it was 220 of something (I planned to run them in parrarel before), guess it'll be back to the drawer...
yep they are silver mica's.
I got that from a friend who dissapointed with his "gainclone"....

Great !

Another OCfer successfully made the 0-100% version.

Another lesson learned for those who is going to build this circuit, when it comes to these dirt cheap components, buy more sets.

hehehe....
I really should have bought more IC before, but I just can resist that tahu gejrot invasion right on the front of the electronics shop's entrance whenever I go there

Hmm.. that is strange.

Now, "WITHOUT" the fan controller, try connect the PWM cable to the Ground wire and power up, this connection is equal to pure 0% duty cycle and tell us how is that fan behaves.

As you read thru this gigantic thread, we all learned that many OCfers encountered so many different behaviours on different PWM fans.

At 0%, some are normal with minimum speed, but not stopped, some like BF's sanyo will be totally stopped, while like Bob's weird one which has the transfer function that it will run at full speed at 0% duty cycle, but strangely at approx 5% dutycycle, it will rotate at the minimum speed. Yes, its quite confusing, but all of these are made and designed for a purpose.

ahh....
my bad, I really forgotten to do that,
now since you mention it I'll try it on my spare controller and fans
thx

I like the 0% = 100% nidec idea, if the voltage controller eats it it'll either short to 5v, short to ground, or fail open.
If it shorts to 5v I get full fan speed and no overheating.
If it shorts to ground I get full fan speed and no overheating.
And if it shorts open I get full fan speed and no overheating.
Beautiful!

It did confuse me when I was trying to figure out if it was PWM by shorting pins to ground though!

well,
just by reading your post makes me got confuse too
hahahaha....
I really had missed the first line, my bad

man that nidec is really something

Well if you look at the last letter of model of the chip it's different... So that's why your having some issues ...

The chip
TI-IM311(P)
ST-IM311

I noticed this before when i purchased the chips from mouser but was not sure if they made a difference.. One chip would be for general and next one would be for precision

ahhh....
I really didn't notice it before, thanks man
just because the ones that came from the TI looks more shinny I think they're better

guess I'll need to short the RPM wire to the ground first to know how the behavior of my fan



thanks a lot for the help guys

 

[mastrdrver]

Did I miss the part where some one used a transistor and a couple of resistors to make a 3-pin fan act like a PWM one?

As soon as I catch up on about 4 days of almost no sleep, maybe I'll get this done so I can fix Zalmans screw up since they can't make their VF3000A/F/N fans PWM........but they can for Zotac.

Idiots.

 

[Brutal-Force]

Hey there OCF. Long time no see. I see this thread took on a life of its own. Congratulations to everyone who has succeeded or at least attempted this mod.

mastrdrver, this thread is more about a pwm controller for a pwm fan. You are looking for a pwm controller for a non-pwm fan, which is a whole different beast all together.

inVain, your controller is very sharp. I tried a couple of times on PCBs, but in the end I preferred the "bug" type that bing showed me. It uses a lot less material, however it requires a very steady hand. I then popped it into a static proof bag. Ideally I wanted to put it into a small box surrounded by silicon, but looking around I could never find an appropriate box without spending 30 bucks on something.

 

[inVain]

hi Brutal,
thx for starting this whole up....

the "bug" type that you and bing mention earlier, I also did it in the past (while I was still an audio DIYer junkies ), that style was also common to be called p2p (pin to pin) connection....

in audio, which in my case the GC (power amplifier based on IC), I found out that this type of soldering turn out to be the best in reproducing the sound because the "direct contact" between components and it's the only way to shorten the signal lengths....

the downside of this style of soldering style; that it is rather difficult to do (at least for me it's difficult , coz I end up breaking few of my precious ICs trying to do this), as you mention, a pair of steady hands is defininetly required . and don't forget those aligator clips are very handy on this type of soldering method...(man I do envy you)

as for the mounting purpose.....
since you've done insulating it, maybe you can drill an unused drive bay to mount the potentiometer and the knob, and then you can connect them up by using some wire to the circuit....
I'm sorry since these pics has nothing to do with the controller, but I think you may got the idea:


(click to enlarge)

as you can see on the most top left picture, I drilled a hole on the chassis to mount the pots....
and to connect them with the main "bug" circuit, I wired it up (you can see it on the middle buttom picture)....
by doin' this you might save up the cost of the building project

 

[Omega Destroyer]

First off, thank you everyone for contributing to this thread. I found it an excellent resource and this PWM controller is exactly what I needed.

TL;DR
Can someone please post a Digikey parts list for the 2-98% controller with a 556 chip? I tried finding some of the parts but for example, there are thousands of 1k resistors etc, and I'm not sure what other specifications are important. I would need everything including the PCB, the fan connectors, and 4 pin molex connectors. I would like to mount everything onto the PCB including the fan headers, molex connectors and poti.
/TL;DR

The fans I want to control will not be connected to a computer at all. They will be run off a power adapter like this (one.
I'm not running massive fans, just some Gelid 1500 rpm "silent" fans. They are just not silent enough for my home theatre. My amp is in a tight TV stand and just wanted to provide some circulation so it doesn't get too hot, but the 1500 rpm fans are too loud. I would like to run them at around 1000 rpm. Ideally it would be nice to have something that sensed the air temp and changed the fan speed accordingly but I think that might be too complicated.

I did a search on digikey but there are so many parts with the same spec and I'm not sure what to get exactly.

 

[inVain]

are you sure you don't wanna try on the 0-100?
it have better performance and realllly smoooooooth duty cycle tunning

frankly speaking, I've lost on digikey too (just by lookin' at their first page)
I used to buy the components from the local retailer, so I think I won't be any help with the digikey part list

but after building several controller, I've found out that type of parts doesn't play a vital role (the value of the components does)....

so you can choose your components based on your layout wise though
and remember what has bing told us in his previous post....
just make sure you buy more set of components as a spare

 

[Omega Destroyer]

I was considering making the 0-100% but in reality I will likely not be changing this fan speed too much once it's set up and the 0-100% version seems more complicated.

I've been looking through digikey for a thermistor that can replace the 100 k potentiometer but none of them seem to have the correct curve for what I need.

Basically, I want to swap the 100 k pot with a thermistor that pretty much shuts the fan off at around 25 degrees and sets it to maximum around 50-60 degrees with almost any kind of response curve between those values.

Since I don't really know how to select thermistors, I'm going by trial and error looking at all their curves to see if any fit the bill.

 

[petteyg359]

Basically, I want to swap the 100 k pot with a thermistor that pretty much shuts the fan off at around 25 degrees and sets it to maximum around 50-60 degrees with almost any kind of response curve between those values.

Since I don't really know how to select thermistors, I'm going by trial and error looking at all their curves to see if any fit the bill.

AFAIK, you need an IC to convert the output from the thermistor to something a digital potentiometer can use.

Somebody please correct me if I'm wrong. I'm not even sure whether those would work together, the data sheets aren't showing up for me.

 

[Bobnova]

An IC in the middle translating for a digital pot is the "correct" way to do it.
I actually have a setup breadboarded right now that can do just that, it reads a temp sensor (or any other analog sensor, really), figures out what it all means, and then tells a digital pot to set Blah resistance, that digital pot is in a single-555 type fan circuit at the moment.

Given some time and some money for said time I could totally make a Thing that does what you want it to Omega.
Really if you're willing to learn how to make Arduino type chips tick (it's not that hard, really) you could do it too.
The fun part is that the curve is reprogramable, or you could run an RGBLED to give the outside world an idea of the current temp or curve, and other such things like that.

 

[Omega Destroyer]

Thanks for the offer Bobnova. Again that seems a bit too complicated for what I wanted. I was hoping to just use the analogue output of the thermistor as an automatic variable resistor but I guess getting one with the exact curve that I want would be impossible to find.
Putting in the IC would bump the costs up too much. I'm just looking for something cheap. I would even be happy with a 7V mod but the damn fan doesn't even run at that voltage.

 

[Bobnova]

An attiny85 microcontroller is what I used, it costs $2 or so
A digital pot is ~80 cents.

The difficulty with the thermistor idea is that a pot is two resistors that vary their resistance in a mirrored fashion. When one has less resistance the other has more. You'd need two perfectly matched thermistors for that, tough to find.

Actually, contemplating the datasheets and such, an atmega mcu given a 10mhz clock can output 20khz PWM directly! I'm going to have to get myself a 10mhz resonator and try this out!
8mhz is a standard option, that gives 15.6khz PWM output, that might actually work just fine, we'll have to wait for Bing on that one.

I may have just complicated my life some more, lol.
(Sidenote: I intend to test the above, but my plate is mighty full right now so it'll be a bit)

 

[Omega Destroyer]

The problem is that my knowledge of building circuits is non-existent. I think can follow very simple schematics but this would be the first time I'm going to build something on a PCB.
Designing a circuit is way beyond my league, which is why I was just planning on following the schematic that Bing provided. That being said, if someone could provide a schematic of attiny85 or atmega circuit, I could try and follow it if it's simple.